This invention relates to a method for minimizing the formation of intermetallic compounds in solder joints used in making electrical interconnections in electronic packaging of semiconductor devices.
In electronic packaging there is an industry-wide problem caused by the tendency of solder joints to form intermetallic compounds with the nickel layer typically used to cover the pads to which the solder adheres. Intermetallic compounds are brittle materials that are sources of microcracks and also serve as paths for crack propagation that lead to the premature failure of solder interconnections. This is a serious reliability issue that limits the growth of the emerging BGA (Ball Grid Array) and CSP (Chip Scale packaging) technologies which are currently considered to be vital to the continued growth of the semiconductor industry. For this reason the reduction of intermetallic growth in solder joints is the subject of substantial RandD efforts. The current approach to minimize intermetallic growth is to place limitations on the degree of thermal exposure of the solder joints. These limitations are often found to be inadequate, as aggressive thermal excursions such as in the solder reflow process cannot be avoided.
In a typical solder joint the solder material wets two pads to make the interconnection. Pads are typically I/O sites in the circuit layout of either chips or printed circuit boards. In the case of the chips the metal used in the circuitry is aluminum, while in the case of pCBs it is copper. New chip metallization technologies are now replacing aluminum with copper. A solder-wetting layer must cover the surface of the pad. Upon melting, the solder adheres to this layer and forms a strong bond, which is necessary to achieve high reliability. Since solder does not adhere to aluminum, and copper oxidizes easily, the pads require a solder-wetting layer that will insure reliable solder bonds. There are a limited number of metals with reliable and consistent solder-wetting properties. The most common metal, used almost universally in the electronics industry, is nickel covered with a thin layer of gold. Gold over copper cannot be used because copper diffuses easily through gold and forms an oxide layer. Accordingly, a conventional copper pad, shown in FIG. 1A on a substrate 1, has a layer of nickel 11 over the copper pad 10, with a layer of gold 12 covering the nickel.
Nickel has a strong tendency to form intermetallic compounds with tin. The leading solder materials used in the electronics industry contain tin. When the solder melts during the reflow operation to form a connection or solder joint (as shown in FIG. 1B), the protective gold film is dissolved in the molten solder almost instantly. The gold is sometimes referred to as the xe2x80x9csacrificial noble metal.xe2x80x9d palladium and platinum may also be used for this protective film. Following the gold dissolution, the solder 15 comes into intimate contact with the nickel layer 11, and formation of a nickel/tin intermetallic layer 14 begins, following a short period of nucleation and growth. Formation of a thin intermetallic layer (preferably less than 1 micron thick) has been found to be necessary to achieve a strong solder bond. However, a further increase in thickness compromises the reliability of the solder joints, especially when relatively small volumes of solder are used such as in flip chip and BGA applications.
Intermetallic growth requires significant mobility of nickel atoms through the intermetallic layer to meet and react with tin to form the intermetallic compound. As the thickness of the intermetallic layer increases, its rate of growth diminishes as a parabolic function of time, due to the increase in the length of the path through which the atoms must diffuse. However, this decrease in growth rate occurs after the intermetallic is relatively thick (several microns).
The conventional approach to limiting growth of the intermetallic layer is to limit the thermal exposure of the solder joint. This is not practical in most applications since the solder has to undergo multiple reflow operations and thermal cycling due to power dissipation on the chip.
In addition, the present inventors have observed that the use of pure metals to make contact with the solder does not provide an effective environment to reduce intermetallic formation. Intermetallic formation and its growth are known to be governed by diffusion of nickel atoms forming the intermetallics. It is generally known that pure nickel is a good solder-wettable material and that it produces relatively thick intermetallic layers when in contact with solder. As noted above, it is desirable to form an intermetallic layer less than 1 micron thick. Accordingly, it is desirable to form an effective diffusion barrier for nickel atoms that slows down the rate of intermetallic growth, so that only a thin intermetallic layer is formed.
The present invention addresses the above-described need for a reduced-thickness intermetallic layer when a solder joint to a pad is formed.
In accordance with the present invention, a method is provided for forming a solder joint to a pad, in which the thickness of the intermetallic layer formed by nickel in contact with solder is significantly reduced. This method includes the steps of providing a first layer of nickel on the pad; providing a second layer of a noble metal (preferably gold) overlying the first layer; and annealing the first layer and the second layer, thereby forming an alloy layer including nickel and the noble metal between the first layer and the second layer. Contacting the second layer with a molten solder (where the solder includes tin) then causes the noble metal to dissolve in the solder and the solder to subsequently wet the alloy layer, to form an intermetallic layer including nickel, the noble metal and tin.
This method can realize important advantages. Most important, it decreases significantly the rate of formation of intermetallics, thus improving the mechanical reliability of the solder joint. This method also eliminates the thermal-exposure constraints currently imposed on solder reflow operations.
According to another aspect of the invention, a method is provided for forming a solder joint to a pad, in which an alloy layer is deposited on the pad; the alloy includes nickel and an alloying element selected from the group consisting of gold, palladium and platinum. A layer of gold is then deposited on the alloy layer to form a nickel-alloy/gold structure on the pad. Contacting this structure with a molten solder (where the solder includes tin) then causes the gold to dissolve in the solder and the solder to subsequently wet the alloy layer, to form an intermetallic layer including nickel, gold and tin. The nickel alloy layer and the gold layer may be deposited by plating methods.
According to a further aspect of the invention, a solder joint structure is provided, in which solder including tin is connected to a pad. This solder joint structure includes a layer of nickel disposed on the pad and an alloy layer overlying the nickel layer. The alloy layer includes nickel and an alloying element selected from the group consisting of gold, palladium and platinum. The structure also includes an intermetallic layer between the solder and the alloy layer, the intermetallic layer including nickel, the alloying element and tin; the intermetallic layer is formed by the solder wetting the alloy layer. The alloy layer may be formed by annealing a layer of nickel and a layer of the alloying element.
According to an additional aspect of the invention, a solder joint structure is provided which includes an alloy layer disposed on the pad and an intermetallic layer between the solder and the alloy layer. The alloy layer may be deposited on the pad by plating. The alloy includes nickel and an alloying element selected from the group consisting of gold, palladium and platinum; the intermetallic layer includes nickel, the alloying element and tin, and is formed by the solder wetting the alloy layer.